轮廓波和非负稀疏编码收缩的毫米波图像恢复

Denoising millimeter wave image using contourlet and sparse coding shrinkage

针对毫米波图像存在的分辨率较低的问题,结合局部非负稀疏编码(non-negativesparse coding,NNSC)算法的自适应高阶统计特性以及轮廓波分解的方向性和能量变化特性,提出了一种新的基于轮廓波和NNSC收缩的毫米波图像恢复方法。NNSC算法是近年来发展起来的模拟人类视觉系统信息处理的有效方法。使用NNSC训练得到的特征基向量和最大似然估计(MLE),能够自适应地确定收缩去噪阈值,并把该收缩技术应用到轮廓波变换域,则能够大大减少毫米波图像中的大量未知噪声,提高毫米波图像的恢复质量。采用无噪自然图像验证基于轮廓波和NNSC收缩的图像恢复方法,实验结果证实了所提出的算法的有效性和实用性,表明该方法能够有效地用于低分辨率图像的恢复。

As for the problem of low resolution existing in millimeter wave（MMW）image,a new algorithm is proposed which combine the adaptive-self high-order statistical property of non-negative sparse coding（NNSC）algorithm and the contourlet′s composing orientation as well as the energy variation.The NNSC algorithm,developed in recent years,can efficiently simulate the information processing of human′s visual system.Using the feature basis vectors and the maximum likelihood estimation（MLE）,the shrinkage denoising threshold can be determinate.Further,using this shrinkage technique in the contourlet′s transform field,the much unknown noise contained in millimeter wave image can be effectively reduced,and the quality of the restored MMW image can be improved.A clear natural image is used to prove the image restoration method based contourlet and NNSC shrinkage.The experimental results also testify the efficiency and usefulness of this image restoration method.This shows that our method can be used in restoring images with low resolution.